Alternate Bristol Engine Development (Mercury & Pegasus)

[MiTasol]

So far I can only find one US and three Canadian Patents and the Canadians spelt Roy Fedden's name wrong on this one. None cover the complete Cosmos Mercury unfortunately

 

[Bretoal2]

So far I can only find one US and three Canadian Patents and the Canadians spelt Roy Fedden's name wrong on this one. None cover the complete Cosmos Mercury unfortunately
View attachment 736673

Mitasol, in the "Advanced search" panel of Espacenet, type year = 1920 - applicant : Cosmos

You will get 14 patents.... and 11 for 1922.

Regards

 

[mack8]

Interesting scenario. Based on what i've read over the years the sleeve valve was a waste of time, especially the time and expense to get it to work. Every other big air powers (and some crucial british engines like Merlin, Griffon) did perfectly well without sleeve valves.

So in an ATL, Bristol could focus on improving the Mercury and Pegasus, and bring forward the Hercules (normal Mercury type cylinders, alson note it's the same size as GR-14K) and Centaurus (twin-Mercury, because i like the japanese style of using the same cylinders) by years. With 100 octane fuel they could easily get 1400-1500 HP from the ALT Hercules and 1800-2000 HP from the alt-Centaurus in 1940. It seems the sleeve valve engines had unacceptably big diameters, if they can keep the ALT-Hercules and Centaurus in the region of 1,3meters like Mercury, that's helpful for drag. Nearest japanese equivalents such as Ha-5 family and Ha-219 had excellent diameters of 1,26 and 1,28 metres respectively, and the GR-14K/N was about 1,30 so why can't Bristol match that?

Dunno if an ALT-Taurus with normal cylinders would be useful for anything, maybe it's two small. I do particularily like the idea of Bristolified Tiger though, which at 32.7 litres was the same size as the japanese Kinsei. They got 1300 HP out of the two-speed Kinsei-50 WITHOUT 100 octane fuel, so you'd think Bristol could easily get the same of such an ALT-Tiger? Kinsei diameter was very good too at 1,22 metres.

 

[loftonhenderson]

Interesting scenario. Based on what i've read over the years the sleeve valve was a waste of time, especially the time and expense to get it to work. Every other big air powers (and some crucial british engines like Merlin, Griffon) did perfectly well without sleeve valves.

So in an ATL, Bristol could focus on improving the Mercury and Pegasus, and bring forward the Hercules (normal Mercury type cylinders, alson note it's the same size as GR-14K) and Centaurus (twin-Mercury, because i like the japanese style of using the same cylinders) by years. With 100 octane fuel they could easily get 1400-1500 HP from the ALT Hercules and 1800-2000 HP from the alt-Centaurus in 1940. It seems the sleeve valve engines had unacceptably big diameters, if they can keep the ALT-Hercules and Centaurus in the region of 1,3meters like Mercury, that's helpful for drag. Nearest japanese equivalents such as Ha-5 family and Ha-219 had excellent diameters of 1,26 and 1,28 metres respectively, and the GR-14K/N was about 1,30 so why can't Bristol match that?

Dunno if an ALT-Taurus with normal cylinders would be useful for anything, maybe it's two small. I do particularily like the idea of Bristolified Tiger though, which at 32.7 litres was the same size as the japanese Kinsei. They got 1300 HP out of the two-speed Kinsei-50 WITHOUT 100 octane fuel, so you'd think Bristol could easily get the same of such an ALT-Tiger? Kinsei diameter was very good too at 1,22 metres.

Very interesting points! Agreed on diameter, given OTL foreign designs were so much narrower. < 50.5" seems plausible.

Also I agree the ALT Taurus wouldn't be worth it if you have the functioning ALT Hercules. An over-bored Pegasus might still make sense, but would make the ALT Taurus even more redundant.

What do you think a Bristolfied A/S Tiger would look like in practice? What changes would they need to make to the OTL Tiger?

 

[MiTasol]

Mitasol, in the "Advanced search" panel of Espacenet, type year = 1920 - applicant : Cosmos

You will get 14 patents.... and 11 for 1922.

Regards

Thanks

 

[mack8]

Very interesting points! Agreed on diameter, given OTL foreign designs were so much narrower. < 50.5" seems plausible.

Also I agree the ALT Taurus wouldn't be worth it if you have the functioning ALT Hercules. An over-bored Pegasus might still make sense, but would make the ALT Taurus even more redundant.

What do you think a Bristolfied A/S Tiger would look like in practice? What changes would they need to make to the OTL Tiger?

There are folks here much more versed into engine stuff, i was just imagining a Tiger redone/ crossbred according to Bristol design practice. Based on the rough HP per litre ratio of the later 1000 HP Pegasus models, this ALT-Tiger would do about 1150 HP, which is still a significant increase from OTL (seems at about 900 HP?). But like i mentioned above the Kinsei-50 was doing 1300 HP from the same capacity, so there's room for improvement for this ALT- Bristol Tiger.

 

[Shortround6]

There are folks here much more versed into engine stuff, i was just imagining a Tiger redone/ crossbred according to Bristol design practice. Based on the rough HP per litre ratio of the later 1000 HP Pegasus models, this ALT-Tiger would do about 1150 HP, which is still a significant increase from OTL (seems at about 900 HP?). But like i mentioned above the Kinsei-50 was doing 1300 HP from the same capacity, so there's room for improvement for this ALT- Bristol Tiger.

The Kinsei engines used a center main-bearing. and that is just for starters.
This was the great flaw in all of the Armstrong Siddeley two row radials, the use of a two bearing crankshaft with the whole center section flopping around like an over-cooked noodle between the front and back bearings. This may have been OK or at least tolerable in the 1920s when most people didn't know any better, most engines did not have superchargers, 70 octane fuel was hot stuff, and the valves and valve springs were going to fail before the crankshaft gave out (mostly).
The early engines were light
The Tiger, with it's two speed supercharger, was about 200lbs lighter than an R-1830.

You can fix the Tiger, you just need.

New crankcase and crankshaft.
New connecting rods and pistons.
New supercharger. New Carb.
New Cylinders with a lot more fins.
New Cylinder heads, larger valves, and a lot more fins.
Repeat, a LOT more fins.
New reduction gear.

Maybe you can keep the magneto's ??

Now a lot of this "stuff" was known to engine designers of the time. The Problem was convincing Sir Siddeley that the engine needed changing, A self taught 'engineer' who relied on hunches. A-S had bewildering number of engine models to suit just about every need and fancy. However this was done using about 3 different cylinder sizes (and a few tweaks) and then combining the cylinders into 3, 5 and 7 cylinder units and using two 5s on one crankshaft to make a 10 cylinder two row instead of figuring out how to make a 9 cylinder radial.
Sir Siddeley was also frugal. They needed a whole new engine and started to late, even if the Deerhound would have worked (doubtful).
A major stumbling block for all British radials was providing enough cooling fins. And providing cylinder heads. The US could provide high quality castings that worked up to a certain point, then they shifted for forgings. The British had to switch to forgings sooner as their casting technique wasn't good enough. You have to be able to build what you design. The British were also somewhat dependent on US machine tools. The British could design tools, they could not build enough of them.

 

[Bretoal2]

The Kinsei engines used a center main-bearing. and that is just for starters.
This was the great flaw in all of the Armstrong Siddeley two row radials, the use of a two bearing crankshaft with the whole center section flopping around like an over-cooked noodle between the front and back bearings. This may have been OK or at least tolerable in the 1920s when most people didn't know any better, most engines did not have superchargers, 70 octane fuel was hot stuff, and the valves and valve springs were going to fail before the crankshaft gave out (mostly).
The early engines were light
The Tiger, with it's two speed supercharger, was about 200lbs lighter than an R-1830.

You can fix the Tiger, you just need.

New crankcase and crankshaft.
New connecting rods and pistons.
New supercharger. New Carb.
New Cylinders with a lot more fins.
New Cylinder heads, larger valves, and a lot more fins.
Repeat, a LOT more fins.
New reduction gear.

Maybe you can keep the magneto's ??

Now a lot of this "stuff" was known to engine designers of the time. The Problem was convincing Sir Siddeley that the engine needed changing, A self taught 'engineer' who relied on hunches. A-S had bewildering number of engine models to suit just about every need and fancy. However this was done using about 3 different cylinder sizes (and a few tweaks) and then combining the cylinders into 3, 5 and 7 cylinder units and using two 5s on one crankshaft to make a 10 cylinder two row instead of figuring out how to make a 9 cylinder radial.
Sir Siddeley was also frugal. They needed a whole new engine and started to late, even if the Deerhound would have worked (doubtful).
A major stumbling block for all British radials was providing enough cooling fins. And providing cylinder heads. The US could provide high quality castings that worked up to a certain point, then they shifted for forgings. The British had to switch to forgings sooner as their casting technique wasn't good enough. You have to be able to build what you design. The British were also somewhat dependent on US machine tools. The British could design tools, they could not build enough of them.

Hi Shortround

First Kinsei models had no center bearing (see pic here - from "

Data Base: Japanese Aircraft Engines

" thread).

And don't forget that, for many mi-1930 engines such as Bramo, Piaggio, Hispano-Suiza, "lighter" meant "unable to whithstand high stress", or weak engine....

 

[MiTasol]

The Kinsei engines used a center main-bearing. and that is just for starters.
This was the great flaw in all of the Armstrong Siddeley two row radials, the use of a two bearing crankshaft with the whole center section flopping around like an over-cooked noodle between the front and back bearings. This may have been OK or at least tolerable in the 1920s when most people didn't know any better, most engines did not have superchargers, 70 octane fuel was hot stuff, and the valves and valve springs were going to fail before the crankshaft gave out (mostly).
The early engines were light
The Tiger, with it's two speed supercharger, was about 200lbs lighter than an R-1830.

You can fix the Tiger, you just need.

New crankcase and crankshaft.
New connecting rods and pistons.
New supercharger. New Carb.
New Cylinders with a lot more fins.
New Cylinder heads, larger valves, and a lot more fins.
Repeat, a LOT more fins.
New reduction gear.

Maybe you can keep the magneto's ??

Now a lot of this "stuff" was known to engine designers of the time. The Problem was convincing Sir Siddeley that the engine needed changing, A self taught 'engineer' who relied on hunches. A-S had bewildering number of engine models to suit just about every need and fancy. However this was done using about 3 different cylinder sizes (and a few tweaks) and then combining the cylinders into 3, 5 and 7 cylinder units and using two 5s on one crankshaft to make a 10 cylinder two row instead of figuring out how to make a 9 cylinder radial.
Sir Siddeley was also frugal. They needed a whole new engine and started to late, even if the Deerhound would have worked (doubtful).
A major stumbling block for all British radials was providing enough cooling fins. And providing cylinder heads. The US could provide high quality castings that worked up to a certain point, then they shifted for forgings. The British had to switch to forgings sooner as their casting technique wasn't good enough. You have to be able to build what you design. The British were also somewhat dependent on US machine tools. The British could design tools, they could not build enough of them.

Agree totally but especially with the highlighted bit - a very graphic depiction.

There were one or two other things that could be kept - but, like the magnetos, they were all vendor components

 

[kool kitty89]

As others have mentioned, using a back to back, dual cam disc, 2-row configuration as America, many Japanese, and Fiat's 2-row radial engines used, but retaining Bristol's 4-valve per cylinder system of the Jupiter, Mercury, and Pegasus, sould've been the way to go if Bristol had stuck to popper valve designs.

However, the main counter-arguments to this aren't related to complexity of using the 4 valve arrangement, but instead usually point to Alfa Romeo's difficulties attempting this layout with the 135. It's not clear if the vibration problems suffered by that engine were inherent to this configuration or more specific to Alfa Romeo's own implementation.

The only other 2-row radial implementation using 4 valves per cylinder I'm aware of is the Nakajima Ha5 and its derivative Ha41, 109, and 219, except I haven't found any actual photos of those to confirm they used the same bristol type layout of the earlier Jupiter based engines Nakajima built. I remember seeing at least one such picture years ago, but I might be mis-remembering. It's also not clear they stuck to 4 valves per cyclinder for those later, more powerful developments and all the Ha5 models themselves are in roughly the same power class as the Mercury and Pegasus (or 14K and 14N).

Additionally, US radial engines eventually worked around the 2 valve per cylinder limitation by adopting hemispherical cylinder heads to allow two large valves to be used, achieving similar advantages to 4 smaller valves.

The only other basic disadvantage of the Jupiter/Mercury/Pegasus layout is the forward-facing exhaust valves, meaning you needed 90 degree bends in the exhaust manifold to point them rearward, hence the exhaust collector ring layout being popular with those engines. Though, hypothetically you could still manage decent ejector style stub exhausts with a bit of effort. OTOH this also means all the exhaust valve sides of the cylinder heads are facing forward for maximum access to cooling airflow while also allowing simpler routing of supercharger and intake manifold ducting with the intake valves all at the rear of the cyclinders.

The Alfa Romeo 135 uses that sort of arrangement with all exhaust valves pointing forward and intake to the rear. This also means it's not just a mirrored back-to-back version of the 125, as the exhaust valves would have to face rearward on the back row (plus it uses a 6.5" stroke like the Mercury, not 7.5" like the Jupiter, Pegasus, and 125).

Assuming Bristol could make it work, focusing on a 2-row, 18 cylinder version of the Mercury might be wise as it'd be smaller in diameter, potentially more useful for fighters, and closer to the weight class of the Hercules than a 2-row Pegasus could be. It should've offered a compelling alternative to the Rolls Royce Vulture while not being too big or heavy to be ruled out on at least some of the designs that historically used the Hercules.

If there was something fundamentally troublesome with using dual cams with the 4 valve per cylinder Bristol system, one other possibility for using existing Mercury and Pegasus cylinders could have been in an air-cooled inline arrangement in a W, X, or H configuration. A W-12 layout, using 3 rows of 4 cylinders (like the Napier Lion) especially an inverted W configuration seems like it might have worked well and been simpler than an X-16 or H16 and avoids the dual crankshafts of the latter. 12 Pegasus cylinders with similar volumetric efficiency to the Pegasus XVII, should've had power output competitive with the early Hercules marks, but significantly smaller in frontal area, and probably a fair bit lighter.

If 4-cylinder rows couldn't be cooled effectively, using an X-12 or H-12 layout with 3 cylinders per row might have been effective.


Aside from that, given the performance of the war-time Perseus and reliability issues of the Taurus, it might have made sense to continue to develop and refine the late marks of Pegasus and Mercury engines for the 900-1200 hp class. A 2-speed supercharged version of the mercury should have had potential and both engines had more potential for higher boost pressures if optimized for 100 octane fuel (albeit the latter wasn't a sure thing in the development timeframe of the late 1930s). The Mercury might have been pushed to higher maximum RPM as well, given the piston speeds achieved with the Pegasus at 2600 RPM would be the same as the Mercury at 3000 RPM and one of the advantages of using 4 valves per cylinder is higher maximum RPM. Plus a switch to pressure carbs rather than float carbs to avoid negative G maneuver problems.

Had the Taurus and Perseus been used on particularly low drag or high speed aircraft, the issue of frontal area compared to the Pegasus and Mercury might have been more substantial, but as it was the issue was more marginal and reliability issues with the Taurus were likely a bigger disadvantage than drag reduction, even compared to the Pegasus. (plus a single row radial can have a tighter cowling than a double row one and still have sufficient cooling, and in service the Taurus ended up needing to use larger cowling than initially intended to help address cooling issues, like the Beaufort switching to cowlings similar to the Blenheim's mercury engines). Practically speaking, the Mercury and Taurus may have ended up having similar drag profiles once actually installed on aircraft, also in the same range as the P&W R-1830, while the Pegasus would be close to that of the Wright R-1820.

Albeit, for war-time production, they may have still wanted to streamline production further, and a more heavily boosted Mercury with 100 octane fuel, 2-speed supercharger, and increased RPM might have been the best compromise to fill all the roles the Perseus, Taurus, and Pegasus would otherwise fill.

For aircraft being developed and introduced in the late pre-war period and with the 87 octane constraint in mind, the existing later models of Pegasus should've fairly easily displaced the Perseus and Taurus where take-off power was critical and the Mercury was insufficient for this, and use the Mercury where the Pegasus's frontal area was unacceptable. And for that matter, a 2-speed supercharged version of the Mercury should've closed that gap further, eliminating any of the Perseus models that had better take-off performance and being a better option than the Pegasus where drag was more critical than power. (both for top speed and long range)

For that matter, 2-speed supercharger gearing in the Mercury should've avoided the need for the Bristol Blenheim to carry its limited supply of 100 octane fuel needed for take-off performance.

 

[Shortround6]

However, the main counter-arguments to this aren't related to complexity of using the 4 valve arrangement, but instead usually point to Alfa Romeo's difficulties attempting this layout with the 135. It's not clear if the vibration problems suffered by that engine were inherent to this configuration or more specific to Alfa Romeo's own implementation.

AS a guess it was the vibration problems inherent to the configuration. P&W spent an awful lot of time sorting out the crankshaft on the R-2800. At the time P&W and built 3 different 9 cylinder engines and 3 different 14 cylinder engines. Apparently it was a lot more difficult in practice to add 2 more cylinders to each row than they thought. Wright had a lot of troubles with the early R-3350. The post B-19 engines had been built with about 2in more space front to back in the crankcase than the early engines. Once they started work on the later series engines after the R-2160 Tornado fiasco they filled up the space in the crankcase with vibration dampers.

This also means it's not just a mirrored back-to-back version of the 125, as the exhaust valves would have to face rearward on the back row (plus it uses a 6.5" stroke like the Mercury, not 7.5" like the Jupiter, Pegasus, and 125).

Most English sources (and they may copy from each other) say a 160mm/6.30in stroke for the Alfa 135.

A W-12 layout, using 3 rows of 4 cylinders (like the Napier Lion) especially an inverted W configuration seems like it might have worked well and been simpler than an X-16 or H16 and avoids the dual crankshafts of the latter. 12 Pegasus cylinders with similar volumetric efficiency to the Pegasus XVII,

The W-12 is down on power from a two row 14 using cylinder of the same size. All of these engines use 146mm bore which restricts the valve size.
A lot of the Pegasus engines changed power ratings due to different gear ratios for the supercharger and since the gear ratio of the supercharger also changes the amount of heat imparted to the charge by the supercharger it affects both the density of the charge and total boost limit.

Aside from that, given the performance of the war-time Perseus and reliability issues of the Taurus, it might have made sense to continue to develop and refine the late marks of Pegasus and Mercury engines for the 900-1200 hp class

Well, that span of power shows and increase of 33%. Not saying it is not possible but look at the changes that Wright had to do to go from the 1000hp R-1820G to the 1100hp R-1829 G-100 to the 1200hp R-1820G-200. Then look at a the 1300/1350hp R-1820 H series engines. Basically they kept the same bore and stroke and changed everything else 2 or 3 times.
from an old book

Now please note that the 208 lbs/sq/in BMEP in the text is for the 1200hp R-1820G-200 engine. and just for the cooling problem alone required much different cylinder heads than the 1100hp version.

Albeit, for war-time production, they may have still wanted to streamline production further, and a more heavily boosted Mercury with 100 octane fuel, 2-speed supercharger, and increased RPM might have been the best compromise to fill all the roles the Perseus, Taurus, and Pegasus would otherwise fill.

They were all too small, even running on 100 octane. See the R-1830 and R-1820 engines. The Taurus was 1550 cu in./ starting at about 85% of the displacement of the American engines means you are depending a lot on rpm. Unfortunately the large American engines have more cooling surface, also look at the weights of the American engines.

For that matter, 2-speed supercharger gearing in the Mercury should've avoided the need for the Bristol Blenheim to carry its limited supply of 100 octane fuel needed for take-off performance.

Now the question is if you can actually use the low gear in a two speed Mercury to good advantage. It will help. but you don't get both advantages added together.
By mid 1940 the fuel situation got to point where supplying the Blenheim with 100 octane fuel was a diminishing problem.

 

[MiTasol]

And the older R-1820s were far different to the 1820G's as well.

Did some very minor work on an early 30s Buhl Stearman and from memory that R-1820 engine was around 700hp. It was so different I had to look at the data plate to find what make and model engine it was.